Project Background: The Veterans Health Administration (VA) outsources care through its Community Care (CC) program, a top VA priority. With the 2014 passage of the Veterans Choice Act came rapid expansion of CC. While expanded CC is meant to enhance choice and access, emerging evidence raises concerns about how VA patients experience it. Early CC implementation has been challenging, e.g., with limited availability of CC providers and fragmented care coordination between VA and CC providers. Patient experiences of CC may have special relevance for women, who use far more CC than men. Furthermore, nearly one in five new women Veteran VA patients stops using, or ?attrits? from, VA care, which is three-fold higher odds of attrition than other women in VA. Our team has learned through our current CREATE study that patient experiences of care can impact attrition: in preliminary analyses, odds of attrition in a fiscal year (FY) 2011 cohort were lower for women who received CC care. However, in this post-expansion era of CC, it is not known whether CC will continue to protect against attrition, whether the attrition rate among women will improve or worsen, how specific types of CC may influence women's decisions to continue in versus leave VA care, or what patient subgroups are at particular risk for CC-related attrition. Project Objectives: The Specific Aims are to (1) model CC and other factors expected to predict attrition from VA; (2) examine the longitudinal attrition trajectory pre/post expansion of CC; (3) characterize the facility-level context of CC, to triangulate with Aims 1 & 2 results and inform Aim 4; and (4) examine women's experiences of care and their perspectives on the relationship between CC and plans for future VA use. Project Methods: Guided by a conceptual model that combines the Andersen Behavioral Model and Consumer Choice Theory, we will conduct a mixed methods study designed to achieve integration at all levels: design, methods, interpretation, and reporting. For Aim 1, we will analyze existing data sources in an FY17 national cohort of new women (N~23,000) and new men (N~200,000) VA primary care patients. For Aim 2, we will model the attrition trajectory in sequential national cohorts (FY07-FY19Q2) of all women Veterans new to VA, and separately, men. For Aim 3, we will conduct semi-structured interviews at two timepoints with clinician/staff key stakeholders (n=30) involved in CC at ten purposively selected sites. For Aim 4, we will conduct semi-structured interviews at two timepoints with new women Veteran primary care patients (n=70) at Aim 3 sites who received a CC referral in FY20. Key stakeholders and women Veterans will be asked to provide recommendations for organizational/policy changes that would improve CC and retention of Veterans in VA care. Significance and Relevance to Veterans' Health: Understanding the potential association between CC experiences and attrition among women Veterans represents a critical topic, not only because women are far more likely than men to be referred to CC and often receive gender-specific services that require heavy care coordination, but also because the number of women Veteran VA patients has nearly tripled over the past 15 years. This study addresses the VA priority area of focusing resources more efficiently, by yielding results relevant to system-level and policy changes: it will point to specific types of CC that are associated with greater attrition risk, specific subgroups who are at more risk, and specific elements of VA delivery systems that interact with CC to increase risk of attrition. Next Steps: If this project meets the intended goals, the next step will be to work with our policy partners to identify targets for change and intervention, to minimize attrition and maximize patient-centered care.